X-ray spectroscopy and variability of the luminous quasar PDS 456

Full text

(1)This work is protected by copyright and other intellectual property rights and duplication or sale of all or part is not permitted, except that material may be duplicated by you for research, private study, criticism/review or educational purposes. Electronic or print copies are for your own personal, noncommercial use and shall not be passed to any other individual. No quotation may be published without proper acknowledgement. For any other use, or to quote extensively from the work, permission must be obtained from the copyright holder/s..

(2)

(3) X-ray spectroscopy and variability of the luminous quasar PDS 456. Gabriele Matzeu. A thesis submitted to Keele University for the Degree of Doctor of Philosophy. Department of Physics, University of Keele. January 2017.

(4) iii. Abstract In this thesis I present contemporary X-ray observations of the “Rosetta Stone” of black hole winds, the luminous quasar PDS 456. I perform a detailed analysis of a recent, long Suzaku campaign in 2013 (of ∼ 1 Ms duration) where the X-ray flux was unusually low. During this campaign PDS 456 displays significant short-term X-ray spectral variability, on time-scales of ∼ 100 ks, due to variable absorbing gas crossing the line-of-sight. By investigating the physical properties of these X-ray absorbers I find that they constitute the same inhomogeneous ultra fast accretion disc-wind (vw ∼ 0.25c), which is characterized by highly ionized gas, relatively close to the SMBH and colder, denser clumpy material located further out. A series of five simultaneous observations of PDS 456 with XMM-Newton & NuSTAR then resolved the P-Cygni like profile at Fe K, confirming that the absorption originates from a fast (vmax = 0.35 ± 0.02c), wide angle (i.e., Ω & 2π) wind, capable of causing significant feedback between the black hole and its host galaxy. Collating all the Suzaku and XMM-Newton & NuSTAR observations from 2007–2014, I show that the wind velocity appears correlated with the X-ray luminosity, which may imply that the wind is radiatively driven. The last part of the project is focused on the broadband analysis of PDS 456 from the 2007 archival Suzaku data (∼ 370 ks in duration), where the quasar was observed in a high-flux state and the unabsorbed AGN primary continuum was revealed. Significant spectral variability is present during the intrinsic X-ray flares that are present in this observation, which are likely driven by fluctuations in a two-component (accretion disc plus corona) continuum. This takes the form of a variable soft X-ray excess (< 1 keV), likely to be the Comptonized tail of the accretion disc emission, as well as a high energy powerlaw. The X-ray emission in PDS 456 appears to originate from two distinct regions:- (i) a warm, optically thick layer of gas blanketing the disc, which is responsible for the soft X-ray excess and (ii) a hot/thin electron corona above the disc, which produces the hard X-ray emission. The latter is typically 10Rg in extent and appears to be Compton cooled by soft X-ray emission..

(5) iv. Acknowledgements I would like to express my ultimate gratitude to my supervisor, Dr. James Reeves who, due to his enormous knowledge and passion for X-ray astronomy, has provided me with indispensable guidance over the last four years. Without his immeasurable support this thesis would not have materialized. His devotion to, and his excitement for science made me “step up my game” and for this I will be for ever grateful. I would also like to thank Dr. Emanuele Nardini for his invaluable support, his scientific knowledge and for helping me keep my ‘sanity’ intact. During this Ph.D I also had the pleasure of working with Dr. Valentina Braito whose technical and scientific advice I am very grateful for. These past four years have certainly not been a solitary journey and I was lucky enough to share mine with Michele Costa whose exceptional computing skills helped me a great deal and I cannot thank him enough. I would also like to thank Dr. Jason Gofford for his reassuring support throughout. Without the immense emotional support, strength and patience provided by the wonderful Dawn Waddell I would not have made it this far in any way shape or form! I would also like to thank Clint Clements for being a great friend who I shared many a night cap with as we chatted science late into the night. During this Ph.D I also lost two beloved members of my family, my uncle Toni Gabriele and my brother-in-law Darren Waddell and they will be always in my thoughts. Lastly (but not least) I would like to thank my parents Tina and Raimondo Matzeu who I have not seen enough of during these four years, however their moral support has always been present and strong..

(6) v. Publications Refereed • Reeves, J. N.; Braito,V. ; Gofford, J.; Sim, S. A.; Behar, E.; Costa, M. T.; Kaspi, S.; Matzeu, G. A.; Miller, L.; O’Brien, P. T.; Turner, T. J.; Ward, M. J.; 2014, ApJ, 780, 45,Variability of the High-velocity Outflow in the Quasar PDS 456. • Gofford, J.; Reeves, J. N.; Braito, V.; Nardini, E.; Costa, M. T.; Matzeu, G. A.; O’Brien, P. T.; Ward, M. J.; Turner, T. J.; Miller, L.; 2014, ApJ, 784, 77, Revealing the Location and Structure of the Accretion Disk Wind in PDS 456. • Nardini, E.; Reeves, J. N.; Gofford, J.; Harrison, F. A.; Risaliti, G.; Braito, V.; Costa, M. T.; Matzeu, G. A., G. A.; Walton, D. J.; Behar, E.; Boggs, S. E.; Christensen, F. E.; Craig, W. W.; Hailey, C. J.;Matt, G.; Miller, J. M.; O’Brien, P. T.; Stern, D.; Turner, T. J.; Ward, M. J.; Science, 2015, 347, 860, Black hole feedback in the luminous quasar PDS 456. • Matzeu, G. A.; Reeves, J. N.; Nardini, E.; Braito, V.; Costa, M. T.; Tombesi, F.; Gofford, J.; MNRAS, 2016, 458, 1311-1329, Short-term spectral variability of the quasar PDS 456 observed in a low flux state. • Matzeu, G. A.; Reeves, J. N.; Nardini, E.; Braito, V.; Costa, M. T.; Tombesi, F.; Gofford, J.; Astronomische Nachrichten, 2016, Vol.337, Issue 4-5, p.495, Broadband short term variability of the quasar PDS 456. • Matzeu, G. A.; Reeves, J. N.; Nardini, E.; Braito, V.; Turner, T. J.; Costa, M. T., MNRAS 2017, 465, 2804-2819, X-ray Flaring in PDS 456 Observed in a High Flux State, Monthly Notices of the Royal Astronomical Society.

(7) vi. Contents Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Taxonomy of AGN and the proposed unified scheme . . 1.1.1 Seyfert galaxies . . . . . . . . . . . . . . . . . . 1.1.2 Narrow-line Seyfert 1 galaxies . . . . . . . . . . 1.1.3 Quasars . . . . . . . . . . . . . . . . . . . . . . 1.1.4 Unification of AGN . . . . . . . . . . . . . . . . 1.2 The central engine . . . . . . . . . . . . . . . . . . . . 1.2.1 The accretion process . . . . . . . . . . . . . . . 1.2.2 Accretion disc structure . . . . . . . . . . . . . 1.3 Radiative processes . . . . . . . . . . . . . . . . . . . . 1.3.1 Comptonization . . . . . . . . . . . . . . . . . . 1.3.2 Photoelectric absorption and line emission . . . 1.3.2.1 Bound-free absorption . . . . . . . . . 1.3.2.2 Bound-bound absorption and emission 1.3.2.3 Photoionized absorption . . . . . . . . 1.3.2.4 Fluorescence . . . . . . . . . . . . . . 1.4 The complex AGN spectrum . . . . . . . . . . . . . . . 1.4.1 Power-law Continuum . . . . . . . . . . . . . . 1.4.2 X-ray Reflection . . . . . . . . . . . . . . . . . . 1.4.3 Fe Kα complex emission . . . . . . . . . . . . . 1.4.4 Soft Excess . . . . . . . . . . . . . . . . . . . . 1.4.5 X-ray absorption . . . . . . . . . . . . . . . . . 1.5 AGN-host feedback and outflows . . . . . . . . . . . . 1.5.1 The importance of winds in context of feedback 1.5.2 Spectral signature of outflows . . . . . . . . . . 1.5.2.1 Warm absorbers . . . . . . . . . . . . 1.5.2.2 High velocity iron K absorbers . . . . . 1.6 Aims and motivations of this thesis . . . . . . . . . . . 2 The luminous radio-quiet quasar PDS 456 . . . . . . 2.1 Discovery of PDS 456 . . . . . . . . . . . . . . . . . . . 2.2 The Spectral Energy Distribution of PDS 456 . . . . . 2.2.1 UV properties of PDS 456 . . . . . . . . . . . . 2.3 The Discovery of the broad Fe K absorption profile . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. iii iv v 1 3 4 6 6 7 10 10 12 13 14 15 15 17 18 19 20 20 22 23 25 26 26 29 30 30 34 38 39 39 44 49 52.

(8) vii. 3. Instrumentation, data analysis, statistics and spectral models . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Current X-ray observatories . . . . . . . . . . . . . . . . . . . . . 3.1.1 Suzaku . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2 XMM-Newton . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3 NuSTAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Data reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 Data processing . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Data screening . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3 Suzaku data reduction . . . . . . . . . . . . . . . . . . . . 3.2.4 Reducing data from XIS detectors . . . . . . . . . . . . . . 3.3 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Spectral fitting . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1.1 Construction of the ‘fluxed’ spectra . . . . . . . . 3.3.2 Probability distribution and fit statistics . . . . . . . . . . 3.3.3 The goodness-of-fit . . . . . . . . . . . . . . . . . . . . . . 3.4 Continuum spectral models . . . . . . . . . . . . . . . . . . . . . 3.4.1 Comptonization model: compTT . . . . . . . . . . . . . . . 3.4.2 Multi-temperature Comptonized disc model: optxagnf . . 3.5 Absorption models . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.1 Neutral partial covering models . . . . . . . . . . . . . . . 3.5.2 xstar photoionization code . . . . . . . . . . . . . . . . 3.5.3 Modelling the Fe K absorption with xstar . . . . . . . . 3.5.4 Mildly ionized partial covering absorption . . . . . . . . . 3.6 Photoionized emission models . . . . . . . . . . . . . . . . . . . . 3.6.1 Modelling photoionized Fe K-shell emission . . . . . . . . . 3.7 Table models used in this thesis . . . . . . . . . . . . . . . . . . . 4 Broadband spectral variability of PDS 456 . . . . . . . . . . . . 4.1 Chapter content and motivation . . . . . . . . . . . . . . . . . . . 4.2 Overview of current X-ray observations of PDS 456 . . . . . . . . 4.3 Broadband spectral analysis gof the Suzaku observations . . . . . 4.3.1 Initial comparison between Suzaku and XMM-Newton & NuSTAR . . . . . . . . . . . . . . . 4.3.2 Modelling the broadband SED . . . . . . . . . . . . . . . . 4.4 The Fe K band modelling: emission and absorption profiles . . . . 4.4.1 X-ray background . . . . . . . . . . . . . . . . . . . . . . . 4.4.2 Input SED for xstar photoionization models . . . . . . . 4.4.3 Photoionization modelling of the Fe K absorption in the Suzaku 2013 observations . . . . . . . . . . . . . . . 4.4.4 Photoionization modelling in Suzaku 2007, 2011 and 2013 .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 57 57 57 59 62 65 65 66 68 69 69 70 71 72 75 75 76 78 80 80 83 84 86 88 88 91 92 92 95 98. . . . . .. . . . . .. . . . . .. 99 102 109 110 113. . . . 114 . . . 115.

(9) viii. 4.5. Results of the XMM-Newton/NuSTAR campaign on PDS 456 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.1 Photoionization modelling of the XMM-Newton & NuSTAR observations . . . . . . . . . . . . . . 4.5.2 Comparison of the Fe K variability between the Suzaku and XMMNewton & NuSTAR observations . . . . . . . . . . . . . . . . . 4.5.3 A recombining absorber and a variable emitter . . . . . . . . . . 4.5.4 Estimating the mass outflow rate . . . . . . . . . . . . . . . . . 4.5.5 What is the wind driving mechanism? . . . . . . . . . . . . . . 4.6 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Short-term X-ray spectral variability of the quasar PDS 456 . . . . 5.1 Chapter content and motivation . . . . . . . . . . . . . . . . . . . . . . 5.2 Time-dependent spectral analysis in Suzaku 2013 . . . . . . . . . . . . 5.3 The iron K short-term absorption variability . . . . . . . . . . . . . . . 5.3.1 Gaussian modelling . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2 xstar modelling . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 What causes the continuum short-term spectral variability? . . . . . . . 5.4.1 Partial covering variability . . . . . . . . . . . . . . . . . . . . . 5.4.2 Intrinsic spectral variability . . . . . . . . . . . . . . . . . . . . 5.5 Fractional variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6 Properties of the partial covering absorber . . . . . . . . . . . . . . . . 5.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.7.1 Properties of the clumpy wind and constraints on the X-ray emitting region . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.7.2 Estimate of the wind radial distance from its Keplerian velocity 5.8 Origin and energetics of the flare . . . . . . . . . . . . . . . . . . . . . 5.8.1 Can the flare drive the outflow ? . . . . . . . . . . . . . . . . . . 5.9 Chapter summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 X-ray flaring in PDS 456 observed in a high-flux state . . . . . . . . 6.1 Introduction and content of this chapter . . . . . . . . . . . . . . . . . 6.2 Data reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 Broadband spectral analysis . . . . . . . . . . . . . . . . . . . . . . . . 6.3.1 Modelling the broadband SED . . . . . . . . . . . . . . . . . . . 6.4 Temporal behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4.1 Description of the light curves and softness ratios . . . . . . . . 6.4.2 Flux-flux analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Time dependent spectral analysis . . . . . . . . . . . . . . . . . . . . . 6.5.1 Partial covering changes . . . . . . . . . . . . . . . . . . . . . . 6.5.2 Intrinsic continuum changes . . . . . . . . . . . . . . . . . . . . 6.5.3 Difference spectrum analysis . . . . . . . . . . . . . . . . . . . .. 119 128 131 132 135 137 142 144 144 146 150 150 150 156 156 160 162 165 169 170 172 173 175 177 179 179 181 182 184 191 191 194 197 198 201 203.

(10) ix. 6.5.4 Fractional variability . . . . . . . . . . . . . . . . . . . . . 6.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.6.1 Variable partial covering . . . . . . . . . . . . . . . . . . . 6.6.2 Variable intrinsic continuum . . . . . . . . . . . . . . . . . 6.6.3 Compton cooling of the corona . . . . . . . . . . . . . . . 6.6.4 Energetics and reprocessing . . . . . . . . . . . . . . . . . 6.7 Chapter summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Conclusions and future work . . . . . . . . . . . . . . . . . . . . . 7.1 The key findings of this thesis and comparison with previous work 7.2 PDS 456 in context of AGN–host-galaxy feedback . . . . . . . . . 7.3 Future prospects for PDS 456 and beyond . . . . . . . . . . . . . . 7.3.1 Disc-wind models . . . . . . . . . . . . . . . . . . . . . . . 7.3.2 Future multiwavelength observations of PDS 456 . . . . . . 7.3.3 The “bare” AGN sample . . . . . . . . . . . . . . . . . . . Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . .. 206 210 210 212 213 214 217 219 219 226 231 231 236 240 255.

(11) x. List of Figures 1.1. Spectral Energy Distribution (SED) of a quasar compared to an elliptical galaxy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 A three dimensional classification for AGN . . . . . . . . . . . . . . . . 1.3 Graphic representation of unified scheme of AGN . . . . . . . . . . . . 1.4 Photoelectric absorption in X-rays with varying column density . . . . 1.5 Plot of increase in ionization with fixed column density . . . . . . . . . 1.6 The complex AGN spectrum . . . . . . . . . . . . . . . . . . . . . . . . 1.7 Plot showing the variable reflection spectra . . . . . . . . . . . . . . . . 1.8 Examples of host galaxy-black hole scaling relations . . . . . . . . . . . 1.9 The warm absorber in IRAS 13349+2438 . . . . . . . . . . . . . . . . . 1.10 The warm absorber in MR 2251–178 . . . . . . . . . . . . . . . . . . . . 1.11 The P-Cygni profile in PG1211+143 . . . . . . . . . . . . . . . . . . . 1.12 Schematic of a stratified accretion disc-wind . . . . . . . . . . . . . . . 2.1 Absolute magnitude and optical-infrared luminosity comparison between PDS 456 and 3C 273 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Dereddened optical-infrared spectra of PDS 456 . . . . . . . . . . . . . 2.3 Radio to X-rays spectral energy distribution (SED) of PDS 456 . . . . . 2.4 HST /STIS UV spectrum of PDS 456 compared with a HST /FOS UV mean QSO spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Comparison between the observed-frame wavelength UV spectrum of PDS 456 with NGC 3783 . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6 Absorption trough detected in the Fe K band with RXTE . . . . . . . . 2.7 The data/model residuals from BeppoSAX observations of PDS 456 . . 2.8 The data/model residuals showing Fe K absorption profile in PDS 456 observed with XMM-Newton . . . . . . . . . . . . . . . . . . . . . . . . 2.9 The residual showing the finer structure of the Fe K absorption profile in PDS 456 observed in 2007 with Suzaku . . . . . . . . . . . . . . . . . . 3.1 Diagram of the Suzaku X-ray satellite . . . . . . . . . . . . . . . . . . . 3.2 Diagram of the XMM-Newton X-ray satellite . . . . . . . . . . . . . . . 3.3 A plot showing the relative instrumental effective areas as a function of energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Diagram of the NuSTAR observatory . . . . . . . . . . . . . . . . . . . 3.5 Flowchart outlining the forward-fitting process . . . . . . . . . . . . . . 3.6 CompTT model simulations in both optically thin and thick plasmas . . 3.7 A Schematic model geometry with corresponding spectra of the optxagnf model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.8 zpcfabs model with variable covering fractions . . . . . . . . . . . . . 3.9 zpcfabs model with variable column density . . . . . . . . . . . . . . .. 2 4 9 16 19 21 24 28 31 32 36 37 41 44 46 50 51 53 54 55 56 58 60 62 63 71 77 79 81 82.

(12) xi. 3.10 Plot showing how the depth Fe K absorption trough increases as the column density increases . . . . . . . . . . . . . . . . . . . . . . . . . . 85 3.11 Plot showing the Fe xxv and Fe xxvi resonance absorption as the ionization increases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 3.12 Plot showing the partially ionized partial covering modelled with xstar 88 3.13 Plot showing the photoionized emission from the xstar additive table 90 4.1 Long-term X-ray flux level of the Suzaku and the simultaneous XMMNewton & NuSTAR observations . . . . . . . . . . . . . . . . . . . . . 98 4.2 Broadband fluxed spectra of the 2007, 2011 and 2013 Suzaku and the simultaneous XMM-Newton and NuSTAR observations in 2013/2014 . 100 4.3 Optical to hard X-ray SED including the OM (ObsE), the three Suzaku 2013 sequences and the NuSTAR (ObsE) over the 1 eV - 50 keV energy range105 4.4 Residuals for the different optxagnf model fits over the 1 eV - 50 keV energy range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 4.5 Data to model ratio of the Fe K band corresponding to 2007, 2011, 2013a, 2013b and 2013c Suzaku observations . . . . . . . . . . . . . . . . . . . 108 4.6 Comparison between the 2013 Suzaku XIS03 background subtracted source spectra and the averaged background spectrum . . . . . . . . . . . . . 111 4.7 Optical to hard X-ray Spectral Energy Distribution (SED) of PDS 456 . 112 4.8 Two dimensional contour plots from the xstar parameters . . . . . . 115 4.9 Unfolded rest-frame spectra of the Fe K profile from the five Suzaku sequences overlaid with the best xstar models . . . . . . . . . . . . . 119 4.10 Broadband fluxed spectra of the 2013/2014 simultaneous XMM-Newton and NuSTAR observations. . . . . . . . . . . . . . . . . . . . . . . . . . 122 4.11 Data/Model ratio showing the persistence of the P-Cygni-like profile . . 123 4.12 XMM-Newton and NuSTAR spectra of PDS 456 from the 2013/14 campaign126 4.13 Correlations between xstar parameters and hard X-ray luminosity . . 139 4.14 Time behaviour between hard X-ray luminosity, ionization state and iron K emission flux in XMM-Newton and NuSTAR spectra of PDS 456 . . . . 140 4.15 Log-log correlations between hard X-ray continuum luminosity and the outflow velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 5.1 Long-term behaviour of the broadband X-ray light curves of all the five Suzaku observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 5.2 light curves and normalized softness ratio (0.5 − 1/2 − 5 keV) of the 2013 Suzaku observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 5.3 Fluxed spectra of the eight slices from the 2013 Suzaku observations . . 149 5.4 Evolution of the Fe K absorption short-term variability . . . . . . . . . 152 5.5 Data/model residuals of each individual eight slices fitted with either variable partial covering or intrinsic spectral variability model . . . . . 161 5.6 X-ray fractional variability from the 2013 Suzaku observations . . . . . 163 5.7 Contour plots of the χ2 against the partial covering redshift parameter 166.

(13) xii. 5.8 5.9 5.10 5.11 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12 6.13 6.14 6.15 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9. xstar modelling of the RGS/MOS spectra of ObsE . . . . . . . . . . 169 Schematic representation of a possible structure and location of the outflow170 Zoom in of the data to model ratio plot of slice G . . . . . . . . . . . . 171 light curves in different energy bands and softness ratio from the 2013a Suzaku sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 Flux spectra of the Suzaku and XMM-Newton & NuSTAR observations 183 The XMM-Newton OM photometric data of PDS 456 taken between 2007 and 2014 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186 Optical to hard X-ray spectral energy distribution (SED) of PDS 456 . 187 Residuals for the different optxagnf model fits . . . . . . . . . . . . . . 188 Data to model ratio, compared to the best-fit optxagnf model . . . . . 190 Plots of light curves extracted in different energy bands . . . . . . . . . 192 Plots of the normalized softness ratios between different energy bands . 193 Flux-Flux plot for PDS 456 between 5–10 keV and 0.5–1 keV . . . . . . 195 Flux-flux plots for PDS 456 showing how the behaviours of the three segments are remarkably distinct . . . . . . . . . . . . . . . . . . . . . 196 Plots of the XIS03 spectra from the Suzaku 2007 observation fitted with the partial covering changes model . . . . . . . . . . . . . . . . . . . . 202 Plots of the XIS03 spectra from the Suzaku 2007 observation . . . . . . 203 The net soft flare and the net hard flare difference spectra . . . . . . . 205 Fractional X-ray variability (Fvar ) from the 2007 Suzaku observation . . 207 Schematic of a possible coronal-heated disc geometry in an accreting system215 Simulation of a warm/soft Comptonizing region with compTT . . . . . . 217 Comparison between PDS 456, APM 08279+5255 and 1H 0707-495 spectra224 Time-averaged colour density map of the line-driven disc-wind . . . . . 225 Comparison between the fast and molecular outflow in IRAS–F11119+3257229 Momentum and energy–driven outflows in AGNs . . . . . . . . . . . . 230 Example of distribution of ionization state of Fe through the Sim et al. (2010) wind model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 Examples of synthetic spectra generated from the disc-wind model of Sim et al. (2008, 2010) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 PDS 456 fitted with the Sim et al. (2010) disc-wind model . . . . . . . 236 Simulated spectra of PDS 456 observed with the ATHENA X-IFU . . . 240 Optical to hard X-ray spectral energy distribution (SED) of TON S180 243.

(14) xiii. List of Tables 2.1 3.1 4.1 4.2 4.3 4.4 4.5 4.6 5.1 5.2 5.3 5.4 6.1 6.2 6.3 6.4 6.5. SED values of PDS 456 . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Summary of the data screening and selection criteria for Suzaku XIS detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Summary of the 2007, 2001 and 2013 PDS 456 observation with Suzaku 94 Summary of the five simultaneous XMM-Newton and NuSTAR observations of PDS 456 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 optxagnf model baseline continuum parameters for the 2013 observation with Suzaku . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 Parameters of the variable Fe K absorption profile modelled with xstar 117 Parameters of the variable Gaussian Fe K emission and absorption profiles from XMM-Newton & NuSTAR . . . . . . . . . . . . . . . . . . . . . . 124 xstar parameters of the wind absorption and emission profiles from XMM-Newton & NuSTAR . . . . . . . . . . . . . . . . . . . . . . . . . 130 Fe K Gaussian absorption profile components for eight Suzaku 2013 slices 151 xstar photoionization model components of the Fe K absorption profile for the eight Suzaku 2013 slices . . . . . . . . . . . . . . . . . . . . . . 155 Variable partial covering model parameters for the eight Suzaku 2013 slices158 Variable intrinsic continuum model parameters for the eight Suzaku 2013 slices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Summary of the Suzaku and XMM-Newton & NuSTAR observations . 181 optxagnf best-fit parameters to the Suzaku 2007 spectrum . . . . . . . 185 Gradients and intercepts evaluated from the BCES linear regression fits 197 Best-fit Partial covering changes model best-fit parameters for the three Suzaku XIS 2007 combined segments . . . . . . . . . . . . . . . . . . . 199 Intrinsic Continuum changes best-fit parameters for Suzaku 2007 . . . . 204.

(15) 1. 1. Introduction. It is now established that most galaxies contain a Super Massive Black Hole (SMBH) at their centre, with typical mass ranging from ∼ 106 − 1010 M

(16) (Salpeter 1964; Lynden-Bell 1969; Kormendy & Richstone 1995; Magorrian et al. 1998; Kormendy & Ho 2013). In most galaxies it is largely the nuclear fusion reactions, that take place at the core of their hosted stars, that is responsible for the majority of their energy output (i.e. through the combined electromagnetic radiation emitted from all stars) typically ∼ 1011 L

(17) (where L

(18) ∼ 3.8 × 1033 erg s−1 ). However a small proportion of these galaxies (approximately 10 − 20 %) release an enormous amount of radiation usually observed across the entire electromagnetic spectrum coming from within their inner regions (typically < 100 pc in size). In these small regions bolometric luminosities of the order of Lbol ∼ 1044 −1048 erg s−1 are produced (e.g., Woo & Urry 2002; Lusso et al. 2010) often outshining the rest of the host galaxy by a factor of 100 or more. Galaxies with these properties are identified as active galaxies whilst their central inner regions, or nuclei, are consequentially named as Active Galactic Nuclei (AGN hereafter). Furthermore it has been observed that the large fraction of radiation emitted in active galaxies does not follow the typical pseudo-blackbody spectra observed in stars but rather the contribution of non-thermal radiation (see § 1.3.1) extended over a broader energy distribution. More specifically, the Spectral Energy Distribution (SED hereafter) of an active galaxy differs considerably from a normal galaxy where its emission ranges over the entire electromagnetic spectrum i.e., from radio to γ-rays as shown in Fig 1.1. Consequently in active galaxies, the vast majority of the emitted bolometric luminosity cannot be supplied by stars or interstellar gas but instead is released through different radiation mechanisms, some of which I will discuss later in § 1.3. For these reasons AGN are considered the most energetically persistent astrophysical objects in the Universe that can be observed (in particular quasars) at extremely high redshifts (e.g., z & 7 Mortlock et al. 2011, corresponding to a time when the universe was only ∼ 6% of its current age). These unique properties allow astronomers to probe the early Universe and subsequently.

(19) 2. investigate the evolution of both normal and active galaxies over cosmic time.. Figure 1.1: The SED of the luminous quasar 3C 273 in comparison to that of a more common elliptical galaxy. It is clear that the quasar emission is extended over the full range of the electromagnetic spectrum (from radio to γ-rays) originating from a range of completely different radiation mechanisms while the radiation from the elliptical galaxy is very narrow as it only corresponds to the superposition of stellar spectra. Figure adapted from Schneider (2015).. In this thesis I present the results of my research which concentrates on the X-ray spectroscopic study of the luminous quasar PDS 456. This is particularly focused on the X-ray energy band from the observations carried out with Suzaku, XMM-Newton and NuSTAR space observatories. In this chapter I will explain how the radiation produced by an AGN can undergo different types of reprocessing before reaching the observer, detected through a modified continuum spectrum (by absorption and/or emission) that can supply important information e.g., through variability, on the physical nature of the re-processor. The main motivation of my work has been to investigate the time-scales at which the X-ray spectral variability occurs in PDS 456, which allowed me to gain a deeper understanding of its physical origin..

(20) 1.1 Taxonomy of AGN and the proposed unified scheme. 3. The structure of this chapter consists of: (§ 1.1) an initial general overview of active galaxies, where I will introduce the various classes of AGN in the context of the proposed unified scheme. I then in (§ 1.2) give an outline of the accretion process and the structure of the widely accepted Shakura-Sunyaev accretion disc (Shakura & Sunyaev 1973) as part of the central engine of an AGN. Next in (§ 1.3), I will introduce the most relevant radiation processes that are important at the X-ray energies and in (§ 1.4) I will give a summary of the typical spectral components that characterize an AGN spectrum. Lastly, the outline of the remaining thesis chapters will be presented in (§ 1.6).. 1.1. Taxonomy of AGN and the proposed unified scheme. There are different physical features that have been observed in AGN such as collimated jet emission together with radio lobes, compact luminous centres, long and short time scale spectral variability, strongly Doppler-broadened emission lines. Thus a broad class of AGN including a large variety of sub-classes have been identified. The principal differentiation that characterizes the AGN phenomenology is based on their degree of radio loudness. Thus the ratio between their radio emission at 5GHz and optical in B-band luminosity was investigated in a sample of 114 objects by Kellermann et al. (1989). Hence in the case when. Lν (5GHz) Lν. ≥ 10 then the object is defined as radio-loud. and if not it is radio-quiet (e.g., Wilkes & Elvis 1987; Reeves & Turner 2000a). An additional division can be made in relation to their optical luminosity. On one hand we have a very luminous subclass of AGN named quasi-stellar radio sources or quasars for short, with absolute (B-band) magnitudes brighter than MB = −23 (Schmidt & Green 1983) and on the other, the less luminous (MB > −23) AGN called Seyfert galaxies. A further categorization, leading to yet more diverse subtypes of AGN, depends on the presence or absence of broad optical Balmer lines (i.e. Hα and Hβ) and hence type I or type II AGN respectively. Using these three major divisions, it is possible to place AGN within the three dimensional classification scheme shown in Fig. 1.2 (Tadhunter 2008)..

(21) 1.1 Taxonomy of AGN and the proposed unified scheme. 4. Figure 1.2: A three dimensional classification for AGN in terms of the three major divisions amongst the various classes of AGN: the presence or absence of broad Balmer lines, optical luminosity and radio loudness . The acronyms used are RLQ - Radio Loud Quasar; BL Lac - BL Lacertae Blazar; OVV - Optical Violent Variables; BLRG - Broad Line Radio Galaxy; WLRG - Weak Line Radio Galaxy; RQQ Radio Quiet Quasar; NLRG - Narrow Line Radio Galaxy; FIR - Far Infrared galaxies; LINER - Low ionization Nuclear Emission-line Region. Figure taken directly from Tadhunter (2008). 1.1.1. Seyfert galaxies. Seyfert galaxies (Seyfert 1943) are considered the most common class of AGN observed in the local Universe (i.e. z < 0.05). Their bolometric luminosity ranges from Lbol ∼ 1043 − 1045 erg s−1 . Due to their relative proximity, most Seyfert galaxies are in fact excellent “cosmic” laboratories as their physical processes have been (and currently are).

(22) 1.1 Taxonomy of AGN and the proposed unified scheme. 5. successfully investigated by means of continuously evolving observational techniques (to be discussed in Chapter 3). Seyfert galaxies are separated into two broad types, which are mainly dependent on their relative line widths together with the presence or absence of narrow forbidden lines and Balmer lines (Khachikian & Weedman 1974). These are Seyfert 1 (broad permitted and narrow forbidden lines; Sy1 hereafter) and Seyfert 2 (narrow forbidden lines only; Sy2 hereafter). Sy1 are characterized by the presence of broad optical permitted lines with FullWidth at Half-Maximum (FWHM) > 1000 km s−1 , such as the prominent hydrogen Balmer Hα, Hβ, Hγ as well as C iv and Mg ii lines, and narrow forbidden lines with FWHM < 1000 km s−1 like those from [O iii] lines, and [N ii] among others (e.g., Bianchi, Maiolino & Risaliti 2012). The broad Balmer lines are generated from dense matter with electron density ne ∼ 109 cm−3 and their line widths are typically observed with FWHM ranging between 103 − 104 km s−1 . In contrast the narrow forbidden lines arise from low density gas (i.e. ne ≈ 103 − 106 cm−3 ), as they are subject to collisional de-excitation at higher densities, with Doppler widths of a few hundreds km s−1 (e.g, Osterbrock 1989; Peterson 1997). The broad line emission originates from a region that is typically located relatively close, typically 2 − 27 light days, to the centre (Kollatschny & Zetzl 2013), also referred to as the Broad Line Region (BLR). On the other hand, the narrow emission lines originate in a region that is more extended from a few pc up to ∼ kpc scales (Hainline et al. 2013), which is referred as the Narrow-Line Region (NLR). In Sy2 objects, sometimes named as ‘narrow-line AGN’, both the forbidden and the permitted emission lines show the same narrow widths and they reveal a much weaker continuum which is frequently dominated by their host galaxy, however they do show strong narrow emission lines, especially from forbidden transitions (Trump et al. 2011). The lack of observed emitted broad lines from Sy2s, indicates that their innermost nuclear gas is obscured along the line of sight by a dusty structure, typically at parsec scales, thought to consist of a toroidal geometry surrounding the central accretion disc (e.g., Antonucci 1993; Urry & Padovani 1995; Veilleux, Goodrich & Hill 1997; Nenkova et al. 2008; Bianchi, Maiolino & Risaliti 2012). However, it was discovered that some Seyfert galaxies are characterized by intermediate variants, such as Seyfert 1.5, 1.8 and.

(23) 1.1 Taxonomy of AGN and the proposed unified scheme. 6. 1.9, where the broad lines are still present but with a smaller strength ratio with respect to the narrow lines than Sy1 (Osterbrock 1981).. 1.1.2. Narrow-line Seyfert 1 galaxies. An important subset of Sy1 galaxies are categorized as narrow line emission Seyfert 1 galaxies (NLSy1 hereafter), where one of their main distinctions is based on their optical spectra. The Hβ is narrow of the order of FWHM . 2000 km s−1 which is comparable to a typical Seyfert 1.9 (Osterbrock & Pogge 1985; Goodrich 1989) and they also show strong Fe ii lines, where the ratio Fe ii/Hβ is twice as strong as other Seyferts. NLSy1 are generally radio quiet and X-ray bright, with strong soft (see § 1.4.4) X-ray excesses (e.g., Laor et al. 1994; Boller, Brandt & Fink 1996; Nardini, Fabian & Walton 2012) and exhibit rapid temporal X-ray variability (e.g., Vaughan et al. 1999, 2011; Lobban et al. 2011; Legg et al. 2012; Giustini et al. 2015). Furthermore, these objects are typically characterized by lower black hole masses than other Seyferts, ranging between 105−7 M

(24) ; however, as their bolometric luminosities are still comparable to Seyferts it is suggested that they are accreting at an appreciable fraction of their Eddington rate, defined in § 1.2.1, (Collin & Kawaguchi 2004).. 1.1.3. Quasars. Quasars were initially identified during a series of large surveys of the sky carried out using radio telescopes in the late 50s and early 60s. Their redshifted Balmer emission lines were soon recognised as one of their main features as well as their high luminosities. The quasar 3C 273 was first identified with a cosmological redshift (Schmidt 1963). Two decades later, with the aid of more sensitive observational instrumentation, it was soon recognized that QSOs are a common phenomenon and hence very numerous in the Universe (Schmidt & Green 1983). Quasars are considered the most luminous subclass of AGN outshining their own host galaxy by a factor of 102 − 103 or more, such that only the core is visible, with.

(25) 1.1 Taxonomy of AGN and the proposed unified scheme. 7. Lbol ' 1045 − 1047 erg s−1 and they can be seen very far across the Universe. It follows that these objects are characterized by a stellar-like point source emission, and for this reason the term Quasi-Stellar Object (QSO) was initially adopted in the mid/late 60s. Even though the terms quasar and QSO have been used as a distinction between radio-loud and radio-quiet AGN respectively, nowadays they equally identify a generic AGN where about 10% are radio-loud (Kellermann et al. 1989) producing strong radio emission, powered by collimated relativistic jet, perpendicular to the plane of the accretion disc (see § §1.2.1 and 1.2.2). An interesting class of AGN are named as Broad Absorption Line Quasars BALQSOs (see Hamann & Sabra 2004 for a detailed review), characterized by outflows detected in the optical/UV band. The characteristic spectral signatures imprinted on BALQSOs are the blueshifted (from few ∼ 1000 − 10000 km s−1 ) broad (FWHM ∼ 104 km s−1 ) absorption line (BAL) troughs due to resonant transitions of ionized metals e.g., C iv, N v, O vi, Si iv, emitted from a fast UV wind. Later on in this thesis I will show that some QSOs present very similar characteristics to the BALQSOs in X-rays, such as the broad blueshifted absorption lines that are detected in the iron K band.. 1.1.4. Unification of AGN. The AGN classification can be considered as a stepping stone in identifying their different patterns in behaviour, whereas the AGN unification can be thought as an attempt to understand their fundamental physical properties based on that classification. The basic idea behind the unification scheme is that all types of AGN are intrinsically the same objects. Thus the standard model describes their structure in terms of a central engine, which consists of a central SMBH and an accretion disc. This system is further surrounded by a dusty, pc scale, toroidal structure which is geometrically axisymmetric (e.g., Krolik & Begelman 1988). It follows that the observational differences are due to orientation effects in respect of the line of sight of the observer relative to the “obscuring” toroidal absorber (e.g., Antonucci 1993; Urry & Padovani 1995; Bianchi, Maiolino & Risaliti 2012). A more recent variation of this model is where the “toroidal” structure is.

(26) 1.1 Taxonomy of AGN and the proposed unified scheme. 8. thought to be intrinsically clumpy (e.g., Elitzur 2012). In addition to the central engine, AGN comprise of two further regions, already mentioned in (§ 1.1.1), the broad line region (BLR) and the narrow line region (NLR). It is important to note that the size of the toroidal absorber is large enough to obscure the BLR but small enough not to obscure the NLR (Kaspi et al. 2005; Bianchi, Maiolino & Risaliti 2012) An overall unification scheme is represented schematically in Fig. 1.3, where the type of object that is observed depends on the extent to which the nuclear region is visible. If, for instance, the viewing angle allows the central engine (together with the BLR) to be directly seen, then it follows that the object in question can be classified as type 1 source (or Sy1) where the optical continuum shows broad and narrow lines emitted from the BLR and NLR respectively. On the other hand, if the viewing angle is such that the nuclear region (including the BLR) is impeded by an obscuring opaque torus, then all that is observed are the reprocessed emissions (i.e. absorbed, scattered and reflected) coming from the torus and the NLR. This results in the optical spectrum (in obscured AGN) being dominated by narrow emission lines, hence classifying the object as type 2 source (or Sy2). As it stands, the unification scheme can provide a first order understanding of the AGN phenomenon, predicting differences in appearance based only on the orientation angles relative to the observer, but not accounting for differences in physical properties of the AGN..

(27) 1.1 Taxonomy of AGN and the proposed unified scheme. low power. high power BL Lac. radio-loud (RL) AGN. 9. FSRQ BLRG, Type I QSO. BLRG. jet. NLRG, Type II QSO. NLRG. reﬂected absorbed. radio-quiet (RQ) AGN. Seyfert 2. d mitte trans red scatte. dusty absorber accretion disc electron plasma black hole broad line region narrow line region. Seyfert 1. Figure 1.3: Schematic representation of our understanding of the AGN phenomenon in the unified scheme. The central engine is a SMBH that is actively accreting matter from an accretion disc. Note that the electron plasma region can be also defined as “corona”. Surrounding the nuclear region is a dusty medium of absorbing material of toroidal geometrical shape and the observer line of sight with respect to it determines if the AGN is a type 1 or type 2 Seyfert. The BLR is responsible for the broad permitted emission lines located at sub-parsec scales. The NLR is further out at kpc scales, responsible for the observed narrow forbidden and permitted lines. The same viewing principle applies for the radio-loud objects with the difference that collimated relativistic jet emission is present. FR-I and FR-II are Fanaroff Riley class I and II galaxy respectively, where FSRQ and BL Lac are Flat Spectrum Radio Quasar and BL Lacertae respectively. Figure taken from Beckmann & Schrader (2012)..

(28) 1.2 The central engine. 1.2. 10. The central engine. In this section, the different physical processes that characterize all different types of AGN will be explored, in particular what is recognised as the AGN main source of energy: the accretion of matter onto a SMBH (e.g., Salpeter 1964; Rees 1984).. 1.2.1. The accretion process. One of the most important and efficient astrophysical processes in the Universe is mass accretion, (accretion hereafter) where a gravitational body grows in mass via accumulating (i.e. accreting) matter from an external reservoir. Matter accreting on to a central massive object (such as a SMBH) with mass MBH forms a disc if its angular momentum J is too large for it to fall in directly into the central object. This dynamical argument suggests that the matter will be orbiting the central massive object in a Keplerian manner into a flattened structure called an accretion disc where its angular momentum at circulation radius Rcirc is: Jk =. p GMBH Rcirc .. (1.1). where G = 6.67 × 10−11 m3 kg−1 s−2 (in SI units) is the universal gravitational constant. The total angular momentum of the system is conserved, however for the matter to fall onto the central compact object, the angular momentum of matter from within the inner region of the disc has to be transported outwards far from the centre. Under these conditions the outer regions of the disc gain angular momentum at the expense of the inner regions allowing matter to fall further inward towards the centre. Shakura & Sunyaev (1973) proposed that the accretion disc has a geometrically thin but optically thick structure where in an AGN, matter is accreted constantly into a SMBH. The basic assumption in the accretion disc scenario is that matter falling in towards a black hole with mass MBH liberates gravitational potential energy EGR . For a test mass m which is moved from ∞ to a distance r from the central body, the change in.

(29) 1.2 The central engine. 11. gravitational potential energy is: ∆EGR (r) = −. GmMBH . r. (1.2). Thus if we consider the rate at which matter M is accreted, (where M˙ =. dM ) dt. and. the virial theorem (i.e., 2∆Ekin + ∆EGR = 0) which implies that (for any particle in a stable orbit) only half of the decrease of the gravitational potential (EGR ) is transmitted into the increase in kinetic energy (Ekin ) of the particle, whilst the other half is radiated away as a result of viscosity in the accretion disc (i.e., ∆Ekin = − 12 ∆EGR ) then the accretion luminosity of the system is: GMBH M˙ 1 dEGR = ≡ η M˙ c2 , (1.3) 2 dt 2Rin where η is the degree of accretion efficiency, M˙ is the accretion rate, Rin is the inner Ldisc = −. radial distance from the black hole and c = 2.998 × 108 m s−1 is the speed of light in free space. As matter continues to fall inwards, it eventually reaches the edge of the innermost stable circular orbit (Risco hereafter) which is strongly dependent on the black hole spin. For a non-rotating (Schwarzschild) black hole this is Risco = 6 Rg , where Rg = GMBH /c2 is the gravitational radii. For a maximally rotating (Kerr) black hole this reduces to Risco = 1.235 Rg . Thus in the Schwarzschild case where Rin = Risco =. 6GMBH c2. gives an. efficiency of η = 1/12; however by taking into account general relativistic effects such as photon capture and space-time curvature, the maximum “radiative” efficiency is η = 0.057 in a Schwarzschild metric (Salpeter 1964) and η = 0.32 for a maximally spinning black hole (Thorne 1974). Furthermore for an accretion rate of M˙ = 10 M

(30) year−1 , which is a reasonable value for the quasar PDS 456 (Torres et al. 1997), which I will discuss in the later chapters, I get a disc luminosity of Ldisc ∼ 5 × 1046 erg s−1 if it is accreting down to Rin = 6 Rg and Ldisc ∼ 2 × 1047 erg s−1 if Rin = 1.235 Rg . The luminosity produced by accretion is fundamentally limited to a critical value where the outward radiation pressure equals the inward pressure generated by the gravitational force exerted on the in-falling matter by the SMBH. This critical value is named as Eddington luminosity LEdd and it is expressed as:.

(31) 1.2 The central engine. LEdd = −27. where mp = 1.673 × 10. 12. 4πGMBH mp c M ' 1.3 × 1038 erg s−1 , σT M

(32) kg is the proton rest mass and σT =. (1.4) 8π 3. . e2 4πε0 me c2. 2. '. 6.65 × 10−29 m2 ≡ 6.65 × 10−25 cm2 (in cgs units) is the Thompson cross-section, where e = 1.602 × 10−19 C is the electron charge, ε0 = 8.854 × 10−12 F m−1 is the permittivity of free space and me = 9.109 × 10−31 kg is the electron rest mass. To sustain the Eddington luminosity a critical mass accretion rate is needed: LEdd , M˙ Edd = ηc2. (1.5). ˙ is the Eddington accretion rate. In the case when M˙ acc > M˙ Edd resulting in where MEdd Lacc > LEdd then the excess accreting material will be blown away due to the radiation pressure exceeding the gravitational pull of the SMBH.. 1.2.2. Accretion disc structure. If the disc proposed by Shakura & Sunyaev (1973) is optically thick, then each annulus with area 2πr∆r, radiates as a blackbody with temperature T (r), thus from the StefanBoltzmann law we have: ∆L = 2 × 2πr∆rσSB T (r)4 =. GMBH M˙ ∆r . 2r2. (1.6). Here in the left hand side, σSB T (r)4 is the flux radiated at a radius r on the disc per unit area and 2 × 2πr∆r is the area of the emitting annulus (top plus bottom because the disc has two sides). The whole term on the right hand side is half of the change in gravitational potential through an annulus ∆r of the accretion disc as the particle flows inward. This yields for T (r):. T (r) =. GMBH M˙ 8πσSB r3. ! 41 .. (1.7).

(33) 1.3 Radiative processes. 13. A more complex derivation accounting for the viscous dissipation and in the limiting case where r >> Rg gives:. T (r) = where Rs =. 3GMBH M˙ 8πσSB Rs3. ! 14 . r Rs. − 34 (1.8). 2GMBH c2. is the Schwarzschild radius. After further algebra which includes substituting for Rs in the brackets and expressing M˙ in terms of the Eddington mass accretion rate M˙ Edd (equation 1.5), this can be expressed more conveniently as (see Peterson 1997, page 37):. 7. T (r) = 3.5 × 10 η. M˙ M˙ Edd. − 41. ! 41 . MBH M

(34). − 14 . r Rs. − 43 K.. (1.9). 3. From this it is clear that T (r) ∝ r− 4 so the temperature increases radially inwards, suggesting that the structure of the disc may be characterized by a series of circular annular segments, each emitting as a blackbody. The overall contribution of these annuli can also be described as a superposition of the different blackbody emission in the accretion disc (i.e., multi-colour disc blackbody emission). Furthermore at a given radius, the overall temperature of the disc increases with mass accretion rate but scales down with increasing black hole mass. Thus for a QSO with a typical black hole mass of MBH ∼ 109 M

(35) accreting with η ∼ 0.1 close to its Eddington rate, this would imply the temperature of the inner disc at r = 10 Rs is T ∼ 105 K (corresponding to ∼ 10 eV). Therefore, the thermal blackbody emission for the innermost accretion disc around a SMBH is expected to peak in the UV band (see Fig. 1.6).. 1.3. Radiative processes. In the above section I described the accretion mechanism that powers the central engine of AGN and earlier in § 1.1, I also discussed how the luminosity in specific energy bands and the width of AGN main optical line emissions is adopted as the principal criteria on.

(36) 1.3 Radiative processes. 14. classifying active galaxies into the sub-groups. In this section I describe the relevant radiative and atomic processes caused by the interaction between photons and particles (atoms, ions and electrons) that are important in the X-ray domain.. 1.3.1. Comptonization. Compton scattering is a physical process where an interaction between an high-energy photon and an free non-relativistic (ve c) electron takes place. In this process, when the (X-ray or γ-ray) incident photons are colliding, a portion of the energy and momentum that is carried is transferred to the electrons, which subsequently recoil from the collision. As a result, the photons are scattered off in a new direction with reduced energy and change in momentum. Thus the larger the change in direction of the photon, the larger the energy transfer towards the electron will be. This relationship can be described, in wavelength space, as: λ0 − λ = ∆λ =. h (1 − cos θ), me c. (1.10). where λ is the wavelength of the incident photon, λ0 is the post-scattering wavelength of the photon, θ is the scattering angle and h = 6.626 × 10−34 J s−1 is the Plank constant. Depending on the kinematics of the incident photons and electrons the outcome of the collision would differ rather drastically. Thus when the electrons are relativistic (ve ∼ c) and have more energy than the incident photons, the Compton process can lead to the opposite effect, which is referred as inverse Compton scattering (or Comptonization). In this physical process the low-energy photons can gain in energy and momentum when propagating through a field of relativistic electrons (e.g. in a corona above the accretion disc). The more photon–electron interactions occur the more energy is given to the photons and as the electrons lose their energy, the process will reach a state of energetic equilibrium at the ‘Compton temperature’ (e.g., Fabian et al. 2015). Thus the Compton up-scattering process can be regarded as a cooling mechanism of the electron field (see § 5.6.2) which follows the Comptonization of less energetic photons into the X-ray regime. As I discussed in the previous section the inner accretion disc temperature.

(37) 1.3 Radiative processes. 15. of a typical AGN would be too low to radiate in the X-ray band, thus Comptonization is considered the principal mechanism that is responsible for the observed X-ray emission in the spectrum. Comptonization models used in this thesis will be discussed in § 3.4.. 1.3.2. Photoelectric absorption and line emission. As I mentioned at the beginning of this chapter, the intrinsic X-ray continuum produced in an AGN can be modified by the presence of absorbing material along the line-of-sight observed as spectral features imprinted in the spectrum. At X-ray energies, the most likely photon-matter interaction is through the photoelectric effect, where an atom absorbs a photon with enough energy to remove its single primary electron, usually from the lowest energy level K (n=1), as its binding energy is exceeded. As a result the atom becomes ionized and the atomic transition of the electron through this process is termed as bound-free transition. On the other hand if the energy of the absorbed photon does not exceed the binding energy, the electron will not be ejected but instead it may transit to a higher energy level via bound-bound transition inducing the atom into an excited state. During both transitions a photon is removed from the continuum observed as an absorption feature (or line) into the spectrum.. 1.3.2.1. Bound-free absorption. The key factor that regulates the amount of photoelectric absorption for a given atom/ion i is the optical depth, which is defined as: τi = σ(E)i Ni. (1.11). where σ(E)i is the photoelectric absorption cross-section due to i atom/ion and Ni is the column density of the element i along the line-of-sight with cross-sectional area of 1 cm2 . The former is essentially defined as the cross-sectional area of the atom/ion presented to the photon, which has a complex energy dependence. In the simplest framework, the absorption cross-section of neutral hydrogen is zero below the ionization threshold energy.

(38) 1.3 Radiative processes. 16. of E = 13.6 eV and increases sharply at Eedge = 13.6 eV (peaking at σ = 6 × 10−18 cm−2 ) and then declines as approximately σ(E) ∝ (E/Eedge )−3 at higher energies. Note that the column is made up by heavier elements as well as hydrogen thus the total photo-electric absorption cross-section can be regarded a sum of abundance weighted cross-sections for each element. In X-rays the dominant photo-electric cross-section arises from the K-shell edges of carbon through to iron, as well, as the cross sections of hydrogen and helium above their respective ionization thresholds.. Figure 1.4: Plot showing how the photoelectric absorption in X-rays increases with column density as log(NH /cm−2 ) = 19, 20, 21, 22, 23, 24. Figure adapted from Done (2010). The transmitted (absorbed) and intrinsic continuum emissions are then related by:. −σeff (E)NH. Fobs (E) = Fintr (E) × e. ; where σeff (E) =. i X. σi (E). ni . nH. (1.12). Fobs is the observed observed spectrum and Fintr is the intrinsic continuum spectrum. Thus the bound-free absorption is a process in which an atom becomes ionized as it loses.

(39) 1.3 Radiative processes. 17. one or more electrons depending the energy of the incoming photon. This can have an important effect on the overall photoelectric cross-section as the unbalance in the nuclear charge of the ion (surplus of protons), leads to a greater (electron-nucleus) binding energy and the remaining electrons are more tightly bound. This results in a shift in the Eedge to higher energies for more highly charged ions as shown schematically in Fig. 1.4.. 1.3.2.2. Bound-bound absorption and emission. When an atom absorbs a photon with an energy below the ionization threshold and given that there are sufficient vacancies in the outer electron orbital shells, it can cause a transition of an electron to one of those shells via a line (bound-bound) transition. This process can only occur if the absorbed photon has an energy very close to the quantised energy separation between the different shells. Furthermore when this transition occurs, its signature can be observed as an atomic absorption line imprinted in the spectrum with characteristic energy depending on both the element, its charge state and the specific electron transition. On the other hand when the excited atom returns back to the ground state, the absorption energy is re-emitted as a single (or more) photon with energy characteristic corresponding to the separation between the two orbital shells, which is observed in the spectrum as an emission line. Thus these spectral line features can be considered as the element’s unique ‘fingerprint’ providing information relating to the components and abundances of these elements within the material that is being observed. The strength of these absorption/emission features is measured through its equivalent width (EW ) which is defined as the width of a rectangular segment of continuum (down to zero intensity) centred at the energy of the line, whose area contains the same photon flux as in the line profile. In X-rays, the EW is often measured in units of eV which can be effectively expressed as: EW =. Line intensity (photons cm−2 s−1 ) Continuum flux (photons cm−2 s−1 eV−1 ). (1.13).

(40) 1.3 Radiative processes. 1.3.2.3. 18. Photoionized absorption. Normally a free electron can also recombine with ions, however if the X-ray radiation is intense enough the probability of encountering a high energy photon is more likely before recombination. This leads to the overall absorption cross-section being dominated by photoionized ions, rather than neutral atoms, making the highly ionized gas more transparent to the continuum radiation. Furthermore σ(E) is also strongly dependent on the ionization state of the material as at increasing ionizations the number of bound electrons are dropping resulting in a decreasing contribution of the lighter elements (as they are completely stripped of all their electrons) to the overall absorption crosssection to the limit where there is no photoelectric absorption at all. Fig. 1.5 shows the flux transmitted through a log(NH /cm−2 ) = 22 column of gas at various ionization stages, ranging from log(ξ/erg cm s−1 ) = 0 − 5, highlighting the decreasing absorption cross-section (and therefore more transmitted flux) at higher ionizations. The ionization state of a gas can provide vital information of various physical properties of the absorber (as I will discuss later in Chapter 4 and 5) and can be obtained through the definition of the ionization parameter ξ derived in Tarter, Tucker & Salpeter (1969): ξ = Lion /nH R2 erg cm s−1 ,. (1.14). where Lion is the typically ionizing source luminosity integrated between 1−1000 Rydberg, nH is the hydrogen number density and R is the distance of the ionizing source from the absorbing clouds. Depending on the number of electrons left in the ion, there is a nomenclature that I will also adopt throughout this thesis. More specifically in highly ionized atoms, when they are left with one or two bound electrons they are refereed as hydrogen like (H-like) or helium like (He-like) respectively as the electron configuration resembles that of the hydrogen or helium. Moreover in terms of the H-like ions (e.g., Fe xxvi) I will discuss in later chapters, I will refer to Lyα as the transition from shell transition n = 1s → 2p (in an absorption line) while Lyβ is from n = 1s → 3p. Furthermore I will refer to Kα and.

(41) 1.3 Radiative processes. 19. Figure 1.5: Plot showing how the increase of the ionization state of a constant log(NH /cm−2 ) = 22 layer of gas alters the transmitted flux. Note that when the ionization state is at its highest, the photoelectric absorption is virtually absent as the gas becomes fully ionized to most elements.. Kβ a shell transition from n = 2 → 1 and n = 3 → 1 (in a fluorescent line emission) respectively.. 1.3.2.4. Fluorescence. Fluorescence can occur when a neutral atom (with all its electrons bound) absorbs a high-energy photon that causes the most inner shell (i.e., K-shell) electron to be removed from it. Subsequently the ion may stabilize by filling the electron ‘hole’ with another electron from a less bound shell via the emission of a ‘fluorescence’ photon. Alternatively in the case where the electron hole in the K-shell is filled by an electron from the L-shell, the energy released in the process can auto-ionize an M-shell (or L-shell) electron. The ionization process is also known as the ‘Auger’ effect. The relative probability between.

(42) 1.4 The complex AGN spectrum. 20. fluorescence emission and ionization, given by the fluorescence yield, is strongly dependent upon the nuclear charge i.e., ∝ ∼ Z 3 (Krause 1979) where Z is the proton number. This suggests that a fluorescence line from an heavier element (e.g., as iron) has a much higher probability to be emitted from a lighter one.. 1.4. The complex AGN spectrum. The different spectral model components that characterize the observed AGN X-ray spectra are now introduced. An overview of all these components is schematically represented in Figure 1.6. It is important to clarify that not all of these components are observable in every X-ray spectrum of AGN and the origin of some of them is still subject to debate.. 1.4.1. Power-law Continuum. The accretion disc around a SMBH with typical temperatures kT ∼ 10 − 100 eV (Shakura & Sunyaev 1973), emits the majority of its thermal blackbody radiation at EUV (Extreme Ultra Violet) wavelengths and thus cannot be responsible (at least directly) for the observed X-ray emission. The currently accepted model invokes non-thermal processes as multiple up-scatterings (or inverse-Compton scattering) of photons, originating from the accretion disc, in a putative “corona” of hot (relativistic) electrons typically of a few hundred keV (Haardt & Maraschi 1991; Haardt & Maraschi 1993). The total contribution of these interactions is modelled, over a relatively narrow energy range, (i.e. 2 − 10 keV) as a law continuum. This is parametrized by the photon flux i.e. F (E) ∝ E −Γ photons cm−2 s−1 keV−1 , where the photon-index values have been typically observed in the Γ ∼ 1.7 − 2.5 range (e.g., Nandra & Pounds 1994; Reeves & Turner 2000a; Porquet et al. 2004b). At higher energies the hard X-ray emission may show a characteristic “roll-over” depending on the temperature of the coronal electrons..

(43) 1.4 The complex AGN spectrum. 21. Figure 1.6: Plot showing a toy model including all the main components that characterize the observed AGN optical/UV to X-ray spectrum of a typical Seyfert galaxy. The total observed spectrum is shown in black. The other components shown are: (1, red) the outer disc component which emits as a (colour temperature corrected) blackbody; (2, green) the soft X-ray excess which may arise from the Comptonised disc photons in the ‘warm’ region of the inner disc; (3, blue) primary X-ray power-law continuum arising from the Comptonized inner disc photons in the hot X-ray corona; (4, cyan) relativistically broadened fluorescence Fe Kα emission line; (5, orange) narrow Fe Kα fluorescence line; (6, magenta) Compton reflection component showing the characteristic ‘Compton’ hump peaking at ∼ 30 keV. The typical spectrum can also show strong atomic features from photo-ionized material at energies E . 2 − 3 keV, also known as the warm absorber, as well as at iron K. All these features are explained in more detail in the text..

(44) 1.4 The complex AGN spectrum. 1.4.2. 22. X-ray Reflection. The observed hard X-ray excess (see Figure 1.6) is thought to originate from reprocessing of “primary” X-ray photons continuum in a Compton-thick structure of relatively cold (T < 106 K) gas present in the nuclear region (Guilbert & Rees 1988; Ferland & Rees 1988; Lightman & White 1988). The physics involved in this process can be modelled as an X-ray continuum isotropically illuminating a slab of cold gas (Zdziarski et al. 1994) or gas in a toroidal geometry (e.g., Yaqoob et al. 2007). Thus in context of the AGN unification schemes, the reflection may arise from scattering off the (slab-like) accretion disc, or off the putative pc-scale torus. The X-ray photons entering the slab, interact with the reflector in different manners. For instance, they can be Compton (down) scattered out of the slab by either free or bound electrons; they can be photoelectrically absorbed, subsequently resulting in fluorescent line emission such as from Fe Kα or Fe Kβ. Due to the energy dependence of photoelectric absorption, most of the incident soft X-ray photons are absorbed whereas hard X-ray photons, on the other hand, are rarely absorbed and hence are Compton down scattered back out of the slab. This interplay between the scattering and absorption cross section gives rise to a broad hump-like shape named “Compton-hump” peaking ∼ 30 keV as shown in Figures 1.6 and 1.7. Reflection per se refers to Compton scattered emission and fluorescent line emission produced via the coronal continuum photons illuminating the surface of the disc (or the torus). Thus the reflection spectrum is characterized by a Compton scattering “hump” together with line emission (see Figure 1.7 left panel). Figure 1.7 (right panel) also shows how the reflection spectrum can be blurred by transverse Doppler shifts from the relativistic motion of gas in the inner disc. Spinning black holes are also referred to as Kerr black holes (Kerr 1963), where the unit of the spin parameter a is characterized by the black hole’s angular momentum J divided by its mass MBH where a = J/MBH = 0.998 is for a maximal spin. Especially when the black hole spin is maximal, the spectrum is gravitationally redshifted to lower energies as the inner most stable circular orbit (isco) reduces to ∼ 1.24 Rg (Thorne 1974), as opposed to 6 Rg for a non-spinning Schwarzschild black hole with a = 0 (see § 1.2.1) and the photons have to “climb” a higher gravitational.

(45) 1.4 The complex AGN spectrum. 23. potential. Both panels of Figure 1.7 provide a general overview of how the ionization parameter ξ = 4πF/ne — where F (erg cm−2 s−1 ) is the flux of the illuminating source and ne (cm−3 ) is the gas electron density — have an effect on the reflected spectra from an optically thick slab. In the low ionization case i.e. log(ξ/erg cm s−1 ) = 1 − 2, the reflection spectra are a combination of a rich and complex set of emission lines superimposed on a strongly absorbed continuum as the photoelectric absorption opacity dominates the electron scattering opacity. As the ionization of the scattering gas increases, i.e. log(ξ/erg cm s−1 ) & 3, all the lighter elements (e.g., O, Ne, Mg, Al, Ar and Ca) and eventually Fe, will become completely ionized and the photoelectric absorption will become negligible while the continuum will resemble the original shape of the illuminating powerlaw (i.e. Γ = 2 in this case).. 1.4.3. Fe Kα complex emission. In the iron K-shell band (5 − 10 keV), the X-ray spectra are characterized by two important emission features, the narrow Fe Kα line and broad Fe Kα line, that can provide fundamental information regarding the AGN central engine, from the inner regions of the accretion disc to the parsec scale molecular torus. The majority of AGN spectra show narrow line emission from the neutral iron at 6.4 keV originating from distant reflection from material such as the torus, BLR or the outer region of the accretion disc (Krolik & Kallman 1987; Nandra 2006) whilst its strength also depends on iron abundance. It was also suggested by Fabian et al. (1989) that Fe Kα emissions originating from the inner region of the accretion disc — close to the central SMBH — are subject to broadening due to transverse Doppler shift and gravitational redshift producing the so called “red-wing” (see Figures 1.6 and 1.7) of emission below the central Fe Kα line energy (Fabian et al. 1989; Laor 1991; Tanaka et al. 1995). The X-ray reflection spectra (Figure 1.7) suggest that the narrow Fe Kα can result from non blurred reflection (left panel) from distant gas, while the broad Fe Kα is the result of blurred reflection from.

(46) 1.4 The complex AGN spectrum. 24. Figure 1.7: Left: Reflection spectra due to the illumination of a uniform slab with varying ionization parameter ξ together with line emission corresponding to various chemical elements. Right: Relativistic blurred reflection spectra; the coloured spectra correspond to a non-rotating Schwarzschild black hole (a = 0), while the black spectra correspond to a maximally spinning Kerr black hole (a = 0.998).. relativistic gas close to the black hole (right panel). Furthermore by observing accretion close to the SMBH, the black hole angular momentum could be determined if highly redshifted emission is measured (e.g., Brenneman & Reynolds 2006; Patrick et al. 2011a, 2012). There have been numerous investigations of Seyfert AGN where an apparent broadening of the Fe K feature has been detected (e.g., Reeves et al. 2004; Miller, Turner & Reeves 2008; Reynolds et al. 2009; Patrick et al. 2011b), however the degree of broadening is still debated. Miller, Turner & Reeves (2008, 2009) have also argued that this spectral feature at iron K could be caused by.

(47) 1.4 The complex AGN spectrum. 25. layers of absorbing gas, instead of “blurred” reflected emission from an accretion disc occurring very close to the SMBH due to the continuum curvature imparted by the warm absorber from soft X-rays up to Fe Kα.. 1.4.4. Soft Excess. A smooth soft X-ray emission component below ∼ 1 keV is commonly observed in unabsorbed AGN (Singh, Garmire & Nousek 1985; Turner & Pounds 1988; Porquet et al. 2004a; Nardini et al. 2011; Nardini, Fabian & Walton 2012), where the power-law component simply fails to account for this extra emission. Studies conducted by Gierliński & Done (2004); Porquet et al. (2004a); Piconcelli et al. (2005) have shown that the soft excess cannot originate directly from the Wien tail (i.e. hard tail) of the accretion disc observed in the UV, arising from the black body emission, as previously thought (e.g., Walter & Fink 1993; Czerny et al. 2003) as this thermal continuum requires disc temperatures far higher than expected. More recently Done et al. (2012) suggested that an increase of the disc effective temperature, caused by Comptonisation of the “seed” photons in a cold (kT < 1 keV) optically thick plasma, may be the reason for the observed excess. However a limitation to this scenario is that the physical model requires electron temperatures of the putative “soft” corona to be virtually constant (∼ 0.1 − 0.2 keV) over several decades in AGN mass (Gierliński & Done 2004). Another viable explanation is that the soft excess is due to reflection from the photoionized surface layers located in the inner region (near the SMBH) of the disc itself, where the extreme relativistic blurring, as seen in Figure 1.7, reduces the (narrow) soft X-ray atomic features into a featureless continuum (Fabian et al. 2002; Crummy et al. 2006). Regardless of the validity of the above interpretations, the physical origin of this component it still an open issue and I will discuss it further in chapter 6..

(48) 1.5 AGN-host feedback and outflows. 1.4.5. 26. X-ray absorption. Having introduced the principal physical processes involved in both emission and absorption, the latter has also an important role when observed in different parts of the X-ray spectrum. One initial degree of absorption is given by the cold dusty interstellar material (ISM) present in our Galaxy which requires to be accounted for, depending on the position of the target relative to the Galactic plane. Furthermore in the context of AGN, it has been established through numerous spectroscopic studies that X-ray absorption, as revealed through the presence of discrete absorption lines below E . 2 − 3keV from cosmically abundant elements such as C, N, O, Ne, Mg, Si, S and Fe, is a characteristic of at least ∼ 50% of AGN in the local Universe (e.g., Blustin et al. 2005; McKernan, Yaqoob & Reynolds 2007). Thus the presence of these absorption lines can be associated with circum-nuclear material located in the vicinity of the AGN. In more recent studies, it has been revealed that X-ray absorption associated with He- and H-like iron at E & 6.7 keV is present in approximately 40% of AGN (Tombesi et al. 2010; Gofford et al. 2013). These absorption lines are usually blueshifted with respect to the AGN rest-frame, strongly suggesting that this material is moving along the line-of-sight towards the observer and hence outflowing. In the next section I will describe in more detail the physical properties of the AGN outflows where I will outline their importance in terms of AGN-host galaxy feedback and their spectral signatures in the X-ray spectrum.. 1.5. AGN-host feedback and outflows. AGN may also play an important role in the evolution of their host galaxies. It is now established that the mass of the SMBH is correlated with different properties of the host galaxy bulge e.g., the stellar velocity dispersion (Fig. 1.8). More specifically, the velocity dispersion corresponds to the standard deviation of stellar rotational velocities relative to the mean. Note that from the virial theorem, the kinetic energy (and thus rotational velocity) is directly related to the potential energy of the system. Thus the.

No results found